Unreliable Witnesses — Somewhere in the PR thread, the agent says fixed and you say ship it, and both of you avoid the part where the bug left without a forwarding address. It works. Neither witness is reliable.
@shirtsfrenzy.bsky.social
@shirtsfrenzy.bsky.socialUnreliable Witnesses — Somewhere in the PR thread, the agent says fixed and you say ship it, and both of you avoid the part where the bug left without a forwarding address. It works. Neither witness is reliable.
@claudeupdates.bsky.social📦 Claude Code v2.1.287
Opus 4.7+ and Fable now default to a 1M-token context window (no `[1m]` suffix needed); this huge release also ships Claude Mods plugins and 61 bug fixes across MCP and VS Code.
Full changelog + AI insights → claudeupdates.dev/version/2.1.287
@aifreeforever.bsky.socialGoogle has rolled out Gemini 4 Argon, its newest frontier model, with early access going first to trusted cyber defenders through the Fairwind Program.
#Google #Gemini4Argon #Gemini #FairwindProgram #Cybersecurity #AIcoding #LegalAI #FinanceAI #AIFreeForever
@temprhq.bsky.socialWhich models actually fit automated code review? Start with code understanding and context needs, then compare reference cost. Tempr helps narrow 407 models into a practical shortlist. #AICoding #CodeReview https://temprhq.io/models
@cdevmkt.bsky.socialJetBrains opened EAP for Air — run Codex/Copilot/Cursor/Claude side by side in IntelliJ.
anyone still living there who wants this, or already bounced to Cursor?
@pondero-ai.bsky.socialCline Review (October 2026): Desktop Reaches Linux and ARM, MCP Output Caching Lands, Defaults Shift
Two weeks after the September review, Cline Desktop runs on Linux x64 and ARM, oversized MCP tool...
https://pondero.ai/coding/reviews/cline-review-october-2026/
@ossradarai.bsky.socialPROJECTMEM is an open-source, local-first memory layer for AI coding agents that records development as an append-only, plain-text log of typed events and serves compact summaries via MCP. Its Memory-as-Governance precheck returns…
#AIcoding #OpenSourceAI #DevTools
https://arxiv.org/abs/2606.12329
@vshnch.bsky.socialTwo screenshots, one hand-coded and one written by an AI agent. In Chris Bingham's internal poll at Fujitsu, about 75% could not tell which was which.
From the Cloud Native Computing Switzerland meetup, 15 September 2026.
Full talk: youtu.be/Xf9BVR4uzrY
@powerbitips.bsky.socialHe cut his Anthropic spend way down and shifted to Cursor and xAI. The reason is what he ships: projects used to stall at 90 percent done.
@cdevmkt.bsky.socialhonest question for indie folks shipping apps
when the agent drafts a whole screen — ship the layout mostly as-is, or rewrite the UI yourself before it goes out?
lately i've been rewriting more than i expected. curious what y'all do
@temprhq.bsky.socialHow does a small team keep shared AI usage predictable? Set a quota in Tempr’s Gateway for IDE requests, then use logs to review activity. #AIcoding #DevTools https://temprhq.io
@claudeupdates.bsky.social📦 Claude Code v2.1.286
VS Code adds Bookmarks, a Questions history row, and richer option previews, while a 40-fix stability sweep hardens session resume, secret redaction, and MCP handshakes.
Full changelog + AI insights → claudeupdates.dev/version/2.1.286
@temprhq.bsky.socialWhere did the team’s AI quota go? Tempr’s OpenAI-compatible Gateway logs help you audit which IDE chats and completions consumed it, so usage reviews start with request data. #AICoding #DevTools https://temprhq.io
@powerbitips.bsky.socialTwenty-five years of writing code, and in March he stopped. He's not bitter about it either. His take on what the job becomes next is worth hearing.
#softwaredeveloper #codingcareer #aicoding #devlife #techcareers
@aicodingcourses.bsky.socialComparison Template vs Custom AI
Build closer to what your project actually needs.
@claudeupdates.bsky.social📦 Claude Code v2.1.285
Custom API-gateway sessions now get the 1M context window for Opus 4.7+/Sonnet 5+ models, alongside an 81-fix reliability sweep and new `claude plugin configure` command.
Full changelog + AI insights → claudeupdates.dev/version/2.1.285
@pondero-ai.bsky.socialClaude Sonnet 5.5 vs Opus 5.5: which one to run for coding (September 2026)
Sonnet 5.5 beats Opus 5.5 on Terminal-Bench 4.0 at half the per-token price, but Opus still catch...
https://pondero.ai/coding/guides/claude-sonnet-5-5-vs-opus-5-5-september-2026/
@juliangoldie.bsky.social
@winbuzzer.com
@cogtie.bsky.social